For the time being it's cheap and simple to use the active xo to divide up the frequencies, and do no more with it.
The DEQ is sophisticated enough to deal with it better than I can right now with analog filters.
It seems conceptually neater to first deal with the drivers and get a flat anechoic on axis response with active xo, and then correct the room. However, from a listening point of view, is there an audible difference? I'm not sure. This is now overlapping into another thread I have started on Ultracurve.
I'm now leaning towards a cheap DIY active analogue xo. Looks like I can in fact do it much cheaper than the PA unit.
I can then look into a future upgrade to the DCX Ultradrive. And when I'm into spending serious money I might then retire that unit to surround speakers and use a higher end unit for the mains, such as Clarity EQ.
KWY,
You must have posted the same time as me as I didn't see your comments. It was actually your review of UC that first got my interest (the cheaper version DSP). For reasons I won't go into that won't be immediately apparent, Ultradrive isn't an option right now. Future upgrade instead. I believe the two will work together very well. At the stage I get into Ultradrive, I may in fact also explore dipoles, but that's another topic.
Thanks for your suggestions regarding the active crossover design. It looks like it's a bit beyond my abilities at the moment. If I intended to use analogue active xo in the long term, I would look into it more. However, I'm going to use this as a step up from my not-so-good passive crossovers, later to be upgraded to digital. Then I will have Ultracurve and Ultradrive, which together will be very powerful. I think they will also make stepping up into a higher end system more achievable, getting into horns and dipoles and higher end drivers than I'm using now.
Perhaps. Given my tastes in bass, the subs seem to be more the limit, however, what I'm looking at is getting as much clarity in the mids as possible. So I want to keep as much bass out of the midbass drivers as possible. The midrange and treble has improved quite a lot since I have actually started using my subs and not using the bass tone controls on the mains! (Yes, I know, bad habits)
I'm quite impressed with SOME of their stuff. BFD, UC, UD and their mic - very good value. I was interested in their power amps until I heard how noisy the fans are - I can't live with them. Instead I have found a good secondhand QSC with a very quiet fan, and more rugged. For their intended use I'd be wary due to reliability issues. I wouldn't want to have things fail on the road. Not that this is a concern for me.
I'm starting to think that a lot of PA equipment is very useful for Hifi. It's a competitive market, and as long as PA stuff is used well, and well chosen, I think it can be very good value. There is also the absense of voodoo which a lot of hifi salesmen seem to feed on. My experience with hifi sales people is that they know a lot more about selling than about hifi.
The xo I'm thinking about had the delay only to the woofer in a 3 way. Now if you have a compression horn tweeter, then a 12" mid, then a sub, I would think you would want the delay to the mid, not the woofer in this case as this is more critical than the sub. This is why I say "crazy."
Thanks for your comments. I always find your advice to be very helpful
You must have posted the same time as me as I didn't see your comments. It was actually your review of UC that first got my interest (the cheaper version DSP). For reasons I won't go into that won't be immediately apparent, Ultradrive isn't an option right now. Future upgrade instead. I believe the two will work together very well. At the stage I get into Ultradrive, I may in fact also explore dipoles, but that's another topic.
Thanks for your suggestions regarding the active crossover design. It looks like it's a bit beyond my abilities at the moment. If I intended to use analogue active xo in the long term, I would look into it more. However, I'm going to use this as a step up from my not-so-good passive crossovers, later to be upgraded to digital. Then I will have Ultracurve and Ultradrive, which together will be very powerful. I think they will also make stepping up into a higher end system more achievable, getting into horns and dipoles and higher end drivers than I'm using now.
Perhaps. Given my tastes in bass, the subs seem to be more the limit, however, what I'm looking at is getting as much clarity in the mids as possible. So I want to keep as much bass out of the midbass drivers as possible. The midrange and treble has improved quite a lot since I have actually started using my subs and not using the bass tone controls on the mains! (Yes, I know, bad habits)
I'm quite impressed with SOME of their stuff. BFD, UC, UD and their mic - very good value. I was interested in their power amps until I heard how noisy the fans are - I can't live with them. Instead I have found a good secondhand QSC with a very quiet fan, and more rugged. For their intended use I'd be wary due to reliability issues. I wouldn't want to have things fail on the road. Not that this is a concern for me.
I'm starting to think that a lot of PA equipment is very useful for Hifi. It's a competitive market, and as long as PA stuff is used well, and well chosen, I think it can be very good value. There is also the absense of voodoo which a lot of hifi salesmen seem to feed on. My experience with hifi sales people is that they know a lot more about selling than about hifi.
The xo I'm thinking about had the delay only to the woofer in a 3 way. Now if you have a compression horn tweeter, then a 12" mid, then a sub, I would think you would want the delay to the mid, not the woofer in this case as this is more critical than the sub. This is why I say "crazy."
Thanks for your comments. I always find your advice to be very helpful
paulspencer said:
Yes, it is I
1 octave on both sides wouldn't be enough, you need a little more before the deviation is below the audibility threshold. A full 2 octaves might be overdoing it a little (-48dB) but not by a wide margin imo.
That is indeed quite appealing. I did some tests with LR8 recently, both in experimental surroundings as well as in a 'realworld' multiway, and concluded for now that it sounds closer to the original than LR4, at least round about 2-3KHz, the average mid-twee xo. I'm still wondering about the phase rotation at the xo point and the extra ringing that comes with high order xo's though. I also think there are better slopes in general, but still, LR8 is definately a good choice and it lets you crossover really low if that is what you want/need.
About the delay: I think there is more to it than KYW is telling. Delaying high frequencies is not trivial. Time delay is achieved by shifting phase for a particular frequency range without changing the frequency component. Because high freq have very short wavelength, you need to add many, many degrees of phase shift for any meaningfull time delay to occur. This means multiple cascading delay circuits, which a manufacturer would rather avoid. With low freq all you need is a single 1ste order allpass in order to delay the low freq several ms, or in distance, several feet.
Cheers,
Hans.
Hans,
Isn't the overlap situation worse with lower order filters? And what about passive crossovers? You say this is needed, but given that it's not possible, what options do I have?
I didn't know LR8 was an option in an active xo. I'd like to hear more about this.
Also, I was of the understanding that when delay is digital, this should not be a problem. Suppose a second identical signal was delayed in time, by means of digital delay - no other changes included. Surely this would do the trick?
Isn't the overlap situation worse with lower order filters? And what about passive crossovers? You say this is needed, but given that it's not possible, what options do I have?
I didn't know LR8 was an option in an active xo. I'd like to hear more about this.
Also, I was of the understanding that when delay is digital, this should not be a problem. Suppose a second identical signal was delayed in time, by means of digital delay - no other changes included. Surely this would do the trick?
Konnichiwa,
I COULD help you design this, I'll see if I can scare up some schamatics that I can adapt or just draw it it out in P-Spice some time this week.
First, can you define the X-overpoint for the Tweeter (this is the critical one) and look at the datasheet FR curves (or your own measurements) for the drivers so you could suggest some basic EQ application.
For the Subwoofer I would suggest you build a serious subwoofer controller, by amalgamating the following schematics:
Parametric And Sub-Woofer Equaliser
As room modes really only need attenuating, you could actually omit most of the original circuit and cut it down to this:
That circuit will statically attenuate by around 6db and have around 6K input Impedance.
A subwoofer X-Over is found here:
Subwoofer variable X-over with Low-End EQ
This allows the extreme LF to be fitered out while boosting the adjecent LF frequencies where your sub might already roll off. If not needed simply omit the Highpass or set it such it protects your woofers suitably.
A variable Phase and switched polarity circuit that works okay for subs is here:
Subwoofer Phase Controller
Voila, a cute and very capable subwoofer controller.
I would recomend it as I found that even with a given FIXED LF cutoff for main speakers the subwoofers often need widely varying settings of X-Over frequency, phase etc to truely integrate well in different rooms/positioning most likely because we have serious LF issues in normal rooms.
Subtractive filters in principle are described here:
Building three-way active speakers powered by chip amps
The added HPF's for the main units I would arrange to form a 120Hz 2nd order LPF, this keeps all serious room mode issues in the domain of the subwoofer, where we have hopefully with three notch filters, variable crossover frequency and variable phase laid enough of a foundation to get them under control.
More later.
Ciao T
paulspencer said:
I COULD help you design this, I'll see if I can scare up some schamatics that I can adapt or just draw it it out in P-Spice some time this week.
First, can you define the X-overpoint for the Tweeter (this is the critical one) and look at the datasheet FR curves (or your own measurements) for the drivers so you could suggest some basic EQ application.
For the Subwoofer I would suggest you build a serious subwoofer controller, by amalgamating the following schematics:
Parametric And Sub-Woofer Equaliser
As room modes really only need attenuating, you could actually omit most of the original circuit and cut it down to this:
An externally hosted image should be here but it was not working when we last tested it.
That circuit will statically attenuate by around 6db and have around 6K input Impedance.
A subwoofer X-Over is found here:
Subwoofer variable X-over with Low-End EQ
An externally hosted image should be here but it was not working when we last tested it.
This allows the extreme LF to be fitered out while boosting the adjecent LF frequencies where your sub might already roll off. If not needed simply omit the Highpass or set it such it protects your woofers suitably.
A variable Phase and switched polarity circuit that works okay for subs is here:
Subwoofer Phase Controller
An externally hosted image should be here but it was not working when we last tested it.
Voila, a cute and very capable subwoofer controller.
I would recomend it as I found that even with a given FIXED LF cutoff for main speakers the subwoofers often need widely varying settings of X-Over frequency, phase etc to truely integrate well in different rooms/positioning most likely because we have serious LF issues in normal rooms.
Subtractive filters in principle are described here:
Building three-way active speakers powered by chip amps
An externally hosted image should be here but it was not working when we last tested it.
The added HPF's for the main units I would arrange to form a 120Hz 2nd order LPF, this keeps all serious room mode issues in the domain of the subwoofer, where we have hopefully with three notch filters, variable crossover frequency and variable phase laid enough of a foundation to get them under control.
More later.
Ciao T
Konnichiwa,
Not if they are correctly engineered, meaning each driver is actually fitted to an acoustic response that gives a desired rolloff INCLUDING the drivers pecularities which then results in X-Overs often with either unequal (1st Order LPF / 3rd Order HPF as used by Shahinian among others) or 2nd order electrical slopes and various acoustical slopes, in case of 2nd order electrical filters the drivers are often shaped to follow acoustically a LR4 function.
If they are correctly designed they tend to give much better results than a "cookbook" active X-over without equalisation of the drivers. Using such and then reshaping the response using equalisers is often done in Pro audio but not really the ideal solution.
Sayonara
paulspencer said:
Not if they are correctly engineered, meaning each driver is actually fitted to an acoustic response that gives a desired rolloff INCLUDING the drivers pecularities which then results in X-Overs often with either unequal (1st Order LPF / 3rd Order HPF as used by Shahinian among others) or 2nd order electrical slopes and various acoustical slopes, in case of 2nd order electrical filters the drivers are often shaped to follow acoustically a LR4 function.
paulspencer said:
If they are correctly designed they tend to give much better results than a "cookbook" active X-over without equalisation of the drivers. Using such and then reshaping the response using equalisers is often done in Pro audio but not really the ideal solution.
Sayonara
Hans L said:
Yes, it is I
1 octave on both sides wouldn't be enough, you need a little more before the deviation is below the audibility threshold. A full 2 octaves might be overdoing it a little (-48dB) but not by a wide margin imo.
That is indeed quite appealing. I did some tests with LR8 recently, both in experimental surroundings as well as in a 'realworld' multiway, and concluded for now that it sounds closer to the original than LR4, at least round about 2-3KHz, the average mid-twee xo. I'm still wondering about the phase rotation at the xo point and the extra ringing that comes with high order xo's though. I also think there are better slopes in general, but still, LR8 is definately a good choice and it lets you crossover really low if that is what you want/need.
About the delay: I think there is more to it than KYW is telling. Delaying high frequencies is not trivial. Time delay is achieved by shifting phase for a particular frequency range without changing the frequency component. Because high freq have very short wavelength, you need to add many, many degrees of phase shift for any meaningfull time delay to occur. This means multiple cascading delay circuits, which a manufacturer would rather avoid. With low freq all you need is a single 1ste order allpass in order to delay the low freq several ms, or in distance, several feet.
Cheers,
Hans.
http://www.woodartistry.com/linkwitzlab/frontiers.htm#F
Yes, you would need more overlap with lower order electronic filters. But keep in mind the basic difference between electronic and acoustic xo slopes: only because of the limitation of 'lesser' active xo's, are you forced to use virtually perfect units. In the real world, you incorporate the units natural slopes into the tailormade xo circuit to arrive exactly at the desired acoustic slope. This might be obvious to you by now, but I thought I should repeat it.paulspencer said:
I'm sure 8th order filters are available in factory xo's, but I'm not aware of specific brands/models. The LR8 has a slope that is the same as cascading two 4th order Butterworth filters (like LR4 is the same as two 2nd order BW). The crossover region where the drivers overlap (which in general is the frequency range where errors occur in summation, on or off-axis) is twice as small as with the LR4 (duh). The LR sums perfectly on axis, but in a typical 2-way you will see a large suckout in the offaxis response at and slightly below the crossover frequency. Assuming you are not listening in an anechoic room, you will hear the off axis response via the reflections in the listening room. The effect is even more noticable with LR2 and very small with LR8.
The dip in the feq response does not necessarily have to be bad, but it is a deviation from the ideal response. If you take into account the much desired BBC dip, it makes sense to use LR2 in a 2-way design crossed around about 2.5-3KHz. So in short: LR2 sounds very relaxed and 'nice', LR4 sounds more precise and LR8 sound virtually flat and a bit 'in the face' compared to the LR2 in the 'typical 2-way'.
I'm sorry, I lost you there and it's getting a bit late so my brain is almost shut down
. What specific trick are you referring to? In general, even with DSP's, the ability to delay sound is not unlimited (at all). Do you mean to use an active xo for deviding freq and a seperate dsp for delay? I wouldn't recommend that. It would be better to stick to a single DSP wich performs pretty much all the tricks you can think of. Or did you mean something different?Cheers,
Hans.
My current tweeter is the Vifa D25AG but I would like to later use the Vifa XT tweeter.
IIRC, drivers should be spaced no further apart centre to centre than the xo frequency. So if the min spacing is 145mm c-c, then the highest xo frequency is 2.4 kHz. This would be my intended xo point, unless someone would demonstrate a higher or lower point would be beneficial.
Regarding the suggested parametric sub eq for modes, I intend to use Ultracurve for this purpose.
I already have a subwoofer controller. It is described here ... 2nd order defeatable rumble filter and variable crossover, one band of parametric eq (not very useful).
An externally hosted image should be here but it was not working when we last tested it.
As I don't consider it ideal, I can modify it, but I was thinking that a 3 way LR active xo would in fact be better.
Do I really need phase control and if so why? How do I know how to set it?
While you suggest a fairly high crossover point, I had intended to cross at or below 80 Hz to allow me to place the subs either side of the listening position - in simulations this gives the smoothest response by far. While ultracurve is very powerful to deal with room modes, I think it's best to first do all you can to get a smooth response.
In this application why would I need a variable crossover control to the subs? It seems strange to me to suggest the crossover point to the subs would be different to the point where the mains are rolled off. Suppose the mains are rolled off at 120 Hz and the subs are crossed at 70 Hz - this would mean a gap!
Part of the reason I wanted to use a fixed crossover point is that it allows me to eq the full range before using the active crossover to divide up the frequencies.
Konnichiwa,
Okay, using AD On-Line filter designer, I come up with the following filter as 3rd order HPF and subtractor....
Interactive Design Tools: OpAmps : Active Filter Synthesis
If your Drivers are the P17WH0 their published response is quite smooth, no EQ is needed.
It includes an 80Hz 1st order HGPF and is pretty tolerant to component variations.
Sure, fine. I would probably save that money in your case and put it towards a DCX. As I am seriously a "Fullrange Driver" guy I got a Digital EQ, in your situation the DCX makes a lot more sense though.
Pretty nice.
Why? Because it is not adjustable and thus practically garantees poor driver integration? Is the main attraction of the LR4 X-Over that you think it NEEDS no adjustment? If you desire that whole shooting match to work well, you MUST adjust it.
You don't need ANYTHING. But if you want to make sure you can integrate subs and main system regardless of positioning having a variable phase is highly usefull.
Measurement and listening, like all else.
Room simulations are notoriously flaky. I suggest you be carefull how much stake you put into them. As a rule, the more you can adjust using "knobs" the more likely you are to be able to get good results when simulations fall down. The less you can adjust the more likely is a requirement of a fundamental change.
Hence the point several have made about getting the DCX, NOT the DEQ.
Because crossovers are ALLWAYS and INVARIABLY acoiustic and at low frequencies the room determines these. In order to get reasonable integration you MUST be able to adjust one X-Over or the other, it's usually easier to adjust the Subwoofer high cutoff.
Only if your speakers are theoretically perfect and in an anechonic chamber. In rality you may require either a gap or overlap and phase adjustment to actually get the in room response smooth.
Which BTW is the wrong way to do it, if you want my opinion.
I suggets making sure that each drivers acoustic output is such that they all integrate well at the listening position.
Sayonara
paulspencer said:
Okay, using AD On-Line filter designer, I come up with the following filter as 3rd order HPF and subtractor....
Interactive Design Tools: OpAmps : Active Filter Synthesis
If your Drivers are the P17WH0 their published response is quite smooth, no EQ is needed.
An externally hosted image should be here but it was not working when we last tested it.
It includes an 80Hz 1st order HGPF and is pretty tolerant to component variations.
paulspencer said:
Sure, fine. I would probably save that money in your case and put it towards a DCX. As I am seriously a "Fullrange Driver" guy I got a Digital EQ, in your situation the DCX makes a lot more sense though.
paulspencer said:
Pretty nice.
paulspencer said:
Why? Because it is not adjustable and thus practically garantees poor driver integration? Is the main attraction of the LR4 X-Over that you think it NEEDS no adjustment? If you desire that whole shooting match to work well, you MUST adjust it.
paulspencer said:
You don't need ANYTHING. But if you want to make sure you can integrate subs and main system regardless of positioning having a variable phase is highly usefull.
paulspencer said:
Measurement and listening, like all else.
paulspencer said:
Room simulations are notoriously flaky. I suggest you be carefull how much stake you put into them. As a rule, the more you can adjust using "knobs" the more likely you are to be able to get good results when simulations fall down. The less you can adjust the more likely is a requirement of a fundamental change.
Hence the point several have made about getting the DCX, NOT the DEQ.
paulspencer said:
Because crossovers are ALLWAYS and INVARIABLY acoiustic and at low frequencies the room determines these. In order to get reasonable integration you MUST be able to adjust one X-Over or the other, it's usually easier to adjust the Subwoofer high cutoff.
paulspencer said:
Only if your speakers are theoretically perfect and in an anechonic chamber. In rality you may require either a gap or overlap and phase adjustment to actually get the in room response smooth.
paulspencer said:
Which BTW is the wrong way to do it, if you want my opinion.
I suggets making sure that each drivers acoustic output is such that they all integrate well at the listening position.
Sayonara
Is there any reason why not to use BOTH DCX and DEQ?
I have the opportunity to get DEQ at a VERY attractive price secondhand. I found it at a very good price at about the same time my family asked me what I wanted for my birthday which is on friday. DCX even at well below list price is much too expensive. So for me right now, it's a question of DEQ or nothing. If I could get DCX at the same price, then I would. However, I'd like to eventually get DCX as well. In the meantime I can experiment with active crossovers.
To get the DCX I'd have to put in quite a bit of my own money, and it then becomes a strange kind of gift.
DCX doesn't look very powerful to me as far as eq goes. It doesn't look like that part will do a good job of dealing with the room, unless perhaps only a few bands of parametric eq are required for the bass.
I have wondered if there is going to be significant signal degradation in using both DEQ and DCX, any thoughts on this?
I have wondered if there is going to be significant signal degradation in using both DEQ and DCX, any thoughts on this?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.